Superman Sentences
Author(s): Ad Neeleman and Kriszta Szendröi
Reviewed work(s):
Source: Linguistic Inquiry, Vol. 35, No. 1 (Winter, 2004), pp. 149-159
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/4179266 .
Accessed: 05/03/2012 08:39
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.
http://www.jstor.org
SQUIBS
AND
DISCUSSION
149
Chomsky, Noam. 1981. Lectures on governmentand binding. Dordrecht:Foris.
Chomsky, Noam. 1995. The MinimalistProgram. Cambridge,Mass.:
MIT Press.
Chomsky, Noam. 2001. Derivation by phase. In Ken Hale: A life in
language, ed. by Michael Kenstowicz, 1-52. Cambridge,
Mass.: MIT Press.
Chomsky,Noam, and HowardLasnik. 1993. The theory of principles
and parameters.In Syntax:An internationalhandbookof contemporary research, ed. by Joachim Jacobs, Amim von
Stechow, Wolfgang Stemefeld, and Theo Vennemann,
506-569. Berlin: Walter de Gruyter.
Franks,Steven, and NorbertHornstein. 1991. Secondarypredication
in Russian and proper govemment of PRO. In Control and
grammar, ed. by Richard K. Larson, Sabine Iatridou,Utpal
Lahiri, and James Higginbotham,1-50. Dordrecht:Kluwer.
Hofmann, Johann B., and Anton Szantyr. 1965. Lateinische Syntax
und Stilistik.Munich:Beck.
Hormstein,Norbert. 1999. Movement and control. Linguistic Inquiry
30:69-96.
Hudson, Richard.2002. Case agreement,PRO and structuresharing.
Ms., University College London.
Martin,Roger. 1996. A minimalisttheory of PRO and control. Doctoral dissertation,University of Connecticut,Storrs.
Martin,Roger. 2001. Null Case andthe distributionof PRO.Linguistic
Inquiry32:141-166.
Sigur6sson, Halld6r A. 1991. Icelandic Case-markedPRO and the
licensing of lexical arguments.NaturalLanguage & Linguistic
Theory9:327-363.
SUPERMAN SENTENCES
Ad Neeleman
UniversityCollege London
Kriszta Szendro'i
UiL OTS, Utrecht University
1 Introduction
Suppose Fathercomes home from work and finds Motherin obvious
distress. Then the following discourse may take place:
(1) Father: What happened?
Mother: You know how I think our children should read
decent books. Well, when I came home, rather
thandoing his homework,[lPJohnnywas [vp reading [DP SUPERMAN] to some kid]].
For many valuable comments, we would like to thank audiences at the
1st Euresco Conference on the Syntax of Normal and ImpairedLanguage;
Logic and Language7; InternationalConferenceon the Structureof Hungarian
6; MIT LingLunch;and the 4th Amsterdam-UtrechtWorkshop,"Meaning and
Intonation."We would furtherlike to thankJocelyn Ballantyne,Danny Fox,
Chris Kennedy, Hans van de Koot, Andrew Nevins, and Craige Roberts for
importantcontributionsto this squib.
150
SQUIBS AND DISCUSSION
The discourse status of Mother's reply is complex, as it contains a
contrastive focus inside a contrastive focus inside an all-focus sentence. First, her answer tells Fatherwhat happenedand thereforethe
bracketedIP must be in focus as a whole. Second, the VP reading
Supermanto some kid is contrastedwith doing his homework.Finally,
the DP Supermanis contrasted with decent books. These three foci
share a single phonological marking,namely, the stress on Superman
(indicatedby small capitals).'
One may have doubts about the reliability of wh-questionslike
Whathappened?as a test for identifyingfocus. Thatdoes not invalidate
the claim that (1) is an example of a focus embeddedin a largerfocus:
the contrastsat the level of the VP and the object are enough to motivate this conclusion. The same point can be made on the basis of the
scope of the focus-sensitive operatorsonly and even, as shown by
Father's continuationin (2) (adaptedfrom Krifka 1991:131).
(2) Father: Despite all our efforts! We even, [only2 read
[KAFKA]2 to our children]l!
"Superman sentences" like Mother's reply have hardly been discussed in the literatureon focus. Apart from Krifka's paper, one of
few relevant works is by Jacobs (1991), whose approachwe briefly
turn to below.
Supermansentences pose a challenge to any syntactictheory of
focus. We will argue, however, that they are more problematicfor
theories of focus based on a feature that requires syntactic licensing
(Horvath1986, Brody 1995, Rizzi 1997) than for theories based on a
mappingbetweenprosodyandmeaning(Chomsky1971, Cinque 1993,
Reinhart1995, Neeleman and Reinhart1998, Zubizarreta1998, Horvath 2000, Zubizarretaand Vergnaud2000, Samek-Lodovici2002).
We first spell out the two approachesto focus in more detail (section
2). We then discuss to what extent they can accommodateSuperman
sentences (sections 3-5).
2 Stress-Based and Feature-Based Approaches to Focus
Feature-basedtheories assume (a) that [ + focus] is interpretedat the
LF interfaceand (b) thatit affects stressassignmentat the PF interface
(Jackendoff 1972). Recent approaches in addition assume (c) that
[+ focus] is checked in syntax in the specifier of a designatedfunctional projection (Horvath 1986, Brody 1990, 1995, Laka 1990, E.
Kiss 1995, Rizzi 1997). Checking is responsible for focus-related
movement in languages like Hungarian,while languages like English
have covert focus movement.Thus, in Nick's reply in (3), a memberof
1Any stress on somekidis secondary. (a) It is perceived as less strong
than the stress on Supernanby native speakers.(b) It is optionaland sensitive
to rhythmiceffects, as is typical of secondarystress: whereas somekidin (1)
can remainwithout stress, if a longer DP appearsas the complementof to, the
secondary stress is obligatory.
SQUIBS AND DISCUSSION
151
the Sex Pistols carries[ + focus]. Accordingly,it moves to the relevant
functionalspecifier at LF and is interpretedthere. At the PF interface,
[+ focus] is responsible for the stress shift from the normal stress
position to the subject. Note that internallyto the subject, stress is as
predicted by the Nuclear Stress Rule, as shown by the fact that A
MEMBER of the Sex Pistols comes here a lot would be an infelicitous
reply in (3).
(3) John: So, which celebritiesdid you say frequentthe Jeremy
Bentham?
Nick: Well, [a member [of the SEX PISTOLS]] comes here
a lot.
Stress-basedtheories assume thatevery utteranceis associated with a
set of potential foci, where members of this set are constituentsthat
contain the main stress.
(4) Any constituentthatcontainsthe main stressof an utterance
can be interpretedas the focus of the utterance.
According to this definition, Nick's reply has at least three potential
foci, listed in (5). (5b) is the actual focus in the discourse in (3).
(5) a.
b.
[DP the SEX PISTOLS]
[DP a member of [DP the SEX PISTOLS]]
C. [IP[DPa memberof [DP the SEX PISTOLS]] comes here a
lot]
In addition, the stress-based theory requires an economy condition
minimizingstressshift (or maximizingthe effects of the NuclearStress
Rule), which accountsfor the infelicity of A MEMBER of the Sex Pistols
comes here a lot as a reply to John's question. This answer has the
focus set in (6).
(6) a.
b.
[N MEMBER]
[DP a MEMBERof [DP the Sex Pistols]]
C [IP[DPa MEMBER of [DP the Sex Pistols]] comes here a
lot]
The contextrequires(5b)/(6b) as the actualfocus. However,(Sb) maximizes the effect of the Nuclear Stress Rule, while (6b) requires an
additionalstressshift withinthe subject.Thus,economyconsiderations
favor the reply in (3). As a rule of thumb, "focus projection" from
a markedstress position tends to be blocked. (As we will show later,
such economy considerationscannot be avoided in theories that take
focus to be a syntactic feature and thereforecannot be used to force
a choice between the stress-basedand feature-basedapproaches.)
The definition of focus set in (4) refers to the main stress of an
utterance,ratherthan to stress in general. The reason for this is that
secondary stress does not support focus. If stress is shifted to the
subject of a transitive sentence, the object usually carries secondary
stress. However, if the object is also contrastivelyfocused, this is not
enough. Then, it has to bear main stress, like the subject.
152
SQUIBS AND DISCUSSION
(7) We have a problembecausepeople don't agree.JOHN wants
to rent a VIDEO, but PETER wants to go to the MOVIES.
One of the attractionsof the stress-basedtheory is that it does away
with the focus featureand relates focus and stress in a direct fashion.
This comes at the cost of having to assume communicationbetween
LF and PF.
Overt focus movement in languages like Hungarianmay at first
sight lend supportto the feature-basedaccount,buton closerinspection
this turns out not to be so. In Hungarian,as in English, foci contain
the main stress. Moreover,the motivationfor the movementcould be
the checking of a feature other than [+ focus]: Horvath (2000:202)
argues that this featureis [ + exhaustiveidentification].Alternatively,
an analysis may be based on the fact that nuclear stress falls on the
leftmost constituent in Hungarian.Leftward focus movement can
thereforebe seen as movementinto the main stressposition (Szendr6i
2003).2
3 SupermanSentencesin the Stress-BasedApproach
Let us now turn to the stress-basedaccount of Mother's reply in (1).
Since stress is on Supermanin this example, the focus set associated
with it is as follows:
(8) a. [DP SUPERMAN]
b. [vp reading [DP SUPERMAN] to some kid]
c. [lPJohnnywas [vp reading[DP SUPERMAN] to some kid]]
It is usually assumedthat in any given context, only one candidateis
selected fromthe focus set as the actualfocus, althoughthereis nothing
in the stress-basedtheory that forces this. In fact, if we do allow an
arbitrarynumberof foci to be selected, the interpretationof Superman
sentences follows directly. The targeted pragmaticinterpretationof
the example at hand actuallyrequiresselection of all threecandidates.
Supermanis contrastedwith decent books, readingSupermanto some
kid is contrastedwith doing his homework,and the entire IP carries
new information.
Recall that, in the stress-basedaccount,economy considerations
tend to block "focus projection" from a markedstress position. For
example, an all-focus sentence normallyhas main stress on the object
(if thereis one). This is because no matterwhere stress falls, the entire
sentence is a member of the focus set. But only the candidatewith
stress on the object maximizes the effects of the Nuclear Stress Rule.
In the example in (1), however, there is "focus projection" even
though stress is shifted from the indirect object to Superman.Yet
2
In their appendixB, Koopmanand Szabolcsi (2000) suggest a possible
approachto Hungarianexamplescomparableto (1) in termsof checkingtheory.
It would take us too far afield to discuss how these can be analyzedin a theory
without focus features.
SQUIBS AND DISCUSSION
153
economy is not violated, because Supermanis itself contrasted(with
decent books). This readingis not availableif the effects of the Nuclear
Stress Rule are maximized, as (9) is associated with the focus set in
(10), which does not contain [DP Superman]
(9) [Ip Johnny was [vp reading Supermanto [DP some KID]]]
(10) a. [DP some KID]
b. [vp reading Supermanto [DP some KID]]
c. [IP Johnnywas [vp readingSupermanto [DP some KID]]]
One may thinkthatin Mother'sanswerin (1) the effects of the Nuclear
StressRule could be maximizedby puttingmain stresson both Superman and some kid. The shifted stress on Supermanis unavoidable,as
just explained,but the focus interpretationof the VP and the IP could
be licensed by the stress assigned by the Nuclear Stress Rule. This is
not what happens, however: the only constituentbearingmain stress
in Mother's answer is Superman.In fact, assigning the same level of
stress to some kid as to Supermanis impossible on the intendedreading of the example.
We propose an account in terms of the economy condition in
(11). Suppose that it is better to have a single stress assigned by the
Nuclear Stress Rule than to have a single shifted stress. Suppose,
furthermore,that it is better to have a single shifted main stress than
to have a shifted main stress as well as main stress in the position
favoredby the Nuclear Stress Rule. Then, since a single shifted stress
suffices in (1), additionalmain stresses are ruled out.
(11) a. Minimize the numberof prosodic peaks (given the targeted interpretation).
b. Minimize stress shift (given the number of prosodic
peaks).
It is of course possible to put focus on both the indirect object and
Superman.This cannot be done by using a single stress, since neither
constituentis contained in the other. Hence, on this reading, and on
this readingonly, the sentence will have to be pronouncedwith main
stress on both constituents.
(12) You know how I want our children to read decent books
and how I think it is importantthat Johnnyplays with kids
his own age. Well, when I came home, ratherthan doing
his homework, [IP Johnnywas [vp reading [DP SUPERMAN]
to [DP some SIXTEEN-YEAR-OLD]]].
More precisely, the focus set associatedwith the bracketedIP in (12)
is as follows:
(13) a.
b.
[DP
some SIXTEEN-YEAR-OLD]
[DP SUPERMAN]
c. [vp reading Supermanto [DP some SIXTEEN-YEAR-OLD]]
d. [IP Johnny was [vp reading Supermanto [DP some SIXTEEN-YEAR-OLD]]]
154
SQUIBS
AND
DISCUSSION
The only interpretationsallowed by (13) that are not given by the
focus sets in (8) and (10) are those in which both the directobject and
the indirect object are focused (there may be additionalwide foci).
But since (12) is less economical than either (1) or (9), it is only
admissibleunderone of these interpretations,as in the context in (12).
Notice that if some sixteen-year-oldin (12) does not carrymain
stress, but secondary stress (as given by the Nuclear Stress Rule), it
is impossible to interpretthis constituentas contrastedwith kids his
own age. This is in line with our earlier conclusion that secondary
stress is not sufficient to supportfocus. For the same reason, it is not
possible to maintain that in the example in (1) it is the (optional)
secondary stress on some kid that licenses VP and IP focus. Rather,
the shifted main stress on Supermanmust allow focus projection.3
We concludethatfoci embeddedin foci receive a naturalexplanation in the stress-basedapproach.Supermansentences are a subcase
of sentences containingmultiple foci, with the crucial difference that
each focus is embeddedin the next. It follows from basic assumptions
of the stress-basedtheory (the notions of focus set and economy) that
all foci are indicatedby a single stress.4
4 Feature-BasedTheorieswithoutChecking
As mentionedabove, the feature-basedtheory of focus that we argue
against makes three assumptions:(a) that [ + focus] is interpretedat
the LF interface,(b) thatit affects stressassignmentat the PF interface,
and (c) that it is checked in syntax in the specifier of a designated
functional projection. This is not to say that we take issue with all
theoriesthatpresupposethe existence of a focus feature.In fact, there
are at least two such theories that seem empiricallycorrect.Crucially,
these theories do not adopt assumption (c). As we show, they are
essentially syntacticencodings of the stress-basedapproach.The two
theories are presentedin (14) and (15).
(14) a. A syntactic feature [+ focus] can be added to a lexical
item when it is selected from the lexicon. This feature
is not copied to dominatingnodes.
b. At the LF interface,any constituentcontaininga lexical
item marked [+ focus] can be interpretedas the focus
of the utterance.
3 The claim thata shifted stress allows focus projectionis hardlynew, but
the cases in the literatureinvolve stress shift as a resultof anaphoricdestressing
(Reinhart1995, Williams 1997, Zubizarreta1998, Schwarzschild1999). Superman sentences are differentfrom these cases: some kid in (1) is not anaphoric.
4 An anonymous reviewer points out a problem with Hungarian,
where
an example parallel to (1) would requireboth movement of Supermanto the
focus position and stress on some kid. It can be argued,however, thatthe stress
on some kid is forced if the constituentis not given. See Szendr6i 2003:sec.
11.2, for discussion.
SQUIBS AND DISCUSSION
155
c. At the PF interface, a terminalmarked [ + focus] must
carry main stress.
(15) a. A syntactic feature [+ focus] can be added to a lexical
item when it is selected from the lexicon. This feature
is obligatorilycopied to dominatingnodes untilit reaches
the root.
b. At the LF interface, any constituent bearing [+ focus]
can be interpretedas the focus of the utterance.
c. At the PF interface, a terminalmarked [ + focus] must
carrymain stress.
Both theories employ a rule that constructs a focus set. In the first
theory, this is (14b), the interpretiverule that defines possible foci at
the LF interface.In the second, the syntacticmechanismof obligatory
percolationand the interpretiverule jointly define the focus set (see
(l5a) and (15b)). Moreover,[ + focus] on the terminalis nothingmore
than a syntactic encoding of main stress.
In order to avoid overgeneration,both theories also need to assume the economy conditionin (11) or somethinglike it. The number
of focus features introducedon terminals has to be minimized, as
explained above. In addition, where possible, [+ focus] has to be
placed on the terminalsingled out by the Nuclear Stress Rule. Thus,
in an all-focus sentence, stress is rightmost, while in a sentence in
which the subject is focused, stress is rightmostwithin the subject.
Although both (14) and (15) are descriptively adequate, these
theories face two problems. First, the assumptionof a rule that adds
[ + focus] to a lexical item when selected for mergeris incompatible
with Inclusiveness(Chomsky1995). This principlestatesthatsyntactic
operationscan only refer to lexical features;but [ + focus] cannot be
a lexical feature.As lexical items arenot inherentlyfocused, the feature
must be inserted after an element has been taken from the lexicon.
Second, the behaviorof [ + focus] is decidedly nonsyntactic.We
know of no other syntactic featuresthat obligatorilypercolate to the
root node or do not percolateat all. According to both (14) and (15),
any category that contains the terminalthat introduces[ + focus] is a
potential focus, and neither distinguishes the actual focus from any
potential foci in syntax. In other words, [+ focus] does not encode
actual focus, but potentialfocus.
If we want to include [+focus] in the set of regular syntactic
featuresthat requirechecking (as stated in assumption(c)), we must
allow for percolation,but not make it obligatory.
5 Checking [+ Focus]
Whatwe take to be the standardfeature-basedtheory(insofaras there
is one) makes the following assumptions:
(16) a. A syntactic feature [+focus] can be added to a lexical
item when it is selected from the lexicon. This feature
may be copied to dominatingnodes.
156
SQUIBS
AND
DISCUSSION
b. There is a designated functional head that attractsthe
largest constituent bearing [+focus] (either overtly or
covertly).5
c. At the LF interface, the largest constituent bearing
[ + focus] is interpretedas the focus of the utterance.
d. At the PF interface, a terminal marked [+ focus] must
carry main stress.
The underlyingmotivationfor this theory is that semanticinterpretation and position at LF are linked in a one-to-one fashion (compare
Rizzi's (1996) Wh-Criterion).It is hence not a coincidence that (16b)
and (16c) both mention "the largest constituentbearing [+ focus]."
As a reminderof how this theory works, consider the answer in
(17). Here, hat introduces[ + focus], which percolatesup to the subject
DP, but no further.As a consequence,this DP is moved to the specifier
of a designatedfunctionalprojection.In English, this presumablyhappens afterSpell-Out(althoughsome analysesassumeovertmovement;
see Kayne 1998). At the LF interface, the man in the hat, being the
largest [ + focus] constituent,is interpretedas focus. At the PF interface, hat, being the [ + focus] terminal,receives main stress.
(17) Susan: Do you think anyone in this bathhouse can play
chess?
Jo:
[THE MAN IN THE HAT] is a serious chess player.
Superman sentences pose a serious problem for theories in which
[+focus] is checked. The reason is simple. In the example in (1),
Supermanmust carry [ +focus], as it carries shifted main stress. As
explained, the largest constituentto which this featurepercolateswill
be interpretedas the actualfocus. But in (1) thereare threeactualfoci:
Superman,reading Supermanto some kid, and the whole sentence.
In order to obtain a focus interpretationon the larger constituents,
[+ focus] has to percolate up to the VP and IP levels. At the same
time, in order to obtain a focal reading of the smaller constituents,
[ + focus] must not percolatebeyond the DP and VP levels. This is a
contradiction.
In the theoriesdiscussedin section4, the syntaxand/orphonology
of an utteranceprovidesa set of possible foci from which one or more
can be selected as the actualfocus at the LF interface.But in a theory
that employs checking to link interpretationand position, the syntax
provides the actual focus, not a set of potential foci. This makes it
impossible for a single [ + focus] (a single stressedterminal)to encode
multiple foci.
An anonymousreviewer suggests a variantof the theory of (16)
that avoids the problems mentioned so far. The proposal maintains
(16a) and (16d) but replaces (16b) with (18) and (16c) with (15b).
5 Weabstract
awayfromthepossibilityof pied-piping
infocusmovement.
It is a problemsharedby the stress-based
andfeature-based
accounts.
SQUIBS
AND
DISCUSSION
157
(18) Thereis a designatedfunctionalhead thatattractssome constituentbearing [ + focus] (either overtly or covertly).
On this view, the interpretationof Supermansentencesrelies on multiple copies of a single [ + focus]. As the reviewerpoints out, this makes
the alternativetheory equivalentto the one in (15), except that it assumes an additionaloperationof featurechecking. But this means that
the motivationbehind the feature-checkingaccount is lost. Since the
constituentmoved to the checking position and the constituentinterpreted as focus need not coincide, there is no longer a one-to-one
association between semantic interpretationand position at LF.
The traditionalview of checking can be maintainedif Superman
in (1) carries three features: [+focus,, +focus2, +focus3]. These
would then percolate to different levels: [+focus1] to the DP level,
[+ focus2] to the VP level, and [+ focus3] to the sentence level. The
crucial question that this approachfaces is how a node can contain
multipleidenticalfeatures.Few cases of duplicatefeatures,if any, are
attestedelsewhere. For example, there are no wh-wordsthatintroduce
two wh-features(forcing a multiple question interpretation).In fact,
Neeleman and Van de Koot (2002) arguethat syntactictheory makes
it impossible for features in nodes to occur more than once.6
One could argue that the three foci in (1) are differentin nature
and hence encoded by differentfeatures.It might be possible to distinguish the IP focus, which is not contrastive,from the DP and VP foci,
which are.But thereis no semantic(or pragmatic)reasonto distinguish
[+ focus1] and [ + focus2].
The only alternativethat we can see is to distinguish the three
focus features on Superman in (1) by claiming that the indices 1, 2,
and 3 have syntactic reality. This is what Jacobs (1991) proposes.
Indexing of features, however, is an ad hoc device, needed only to
deal with multipleembeddedfoci. Second, it is conceptuallyquestionable, since it amounts to assigning features to features. Third, once
admitted,it will overgenerate-for instance,it would allow wh-words
to carry multiple wh-features.Fourth, indexing of focus features is
essentially a means of introducingthe notion of focus set into the
feature-basedtheory,because a stressedterminalcan potentiallycarry
an arbitrarynumberof focus features that can potentially be copied
to any dominatingnode.
Even if we allow indexed focus features, there is an additional
issue. In Mother'sanswerin (1), why shouldall threefocus featuresbe
introducedin the same terminal?In otherwords, how can we exclude
a representationin which [+ focus2, + focus3] are generatedon the
rightmost terminal, so that the bracketedsentence in (12) could be
pragmaticallyequivalent to (1)? As far as we can see, this can only
6
Neeleman and Van de Koot (2002:sec. 5) arguethat duplicatefunctions
can be licensed in certaincircumstancesby the constructionof a grid. But the
conditions for this are not met here.
158
SQUIBS
AND
DISCUSSION
be achieved if in additionto the assumptionsin (16) and the syntactic
indexing of features,an economy conditionlike (11) is adopted.Note
that this condition will have to refer to phonological properties of
utterances,as (11) does, since there is no reason why a clusteringof
identical syntactic features in a single terminal would be preferred
over the occurrenceof those featuresin differentterminals.
We conclude,then,thatin orderto deal with Supermansentences,
the checking theory of focus must make essentially the same assumptions as the stress-basedaccount, and more.
References
Brody, Michael. 1990. Some remarkson the focus field in Hungarian.
In UCL workingpapers in linguistics 2, ed. by John Harris,
201-225. London:UniversityCollege London,Departmentof
Phonetics and Linguistics.
Brody, Michael. 1995. Focus and checking theory. In Approachesto
Hungarian 5, ed. by Istvan Kenesei, 31-43. Szeged: JATE.
Chomsky,Noam. 1971. Deep structure,surfacestructureand semantic
interpretation.In Semantics:An interdisciplinaryreader, ed.
by Danny D. Steinberg and Leon A. Jakobovits, 183-216.
Cambridge:CambridgeUniversity Press.
Chomsky,Noam. 1995. TheMinimalistProgram. Cambridge,Mass.:
MIT Press.
Cinque, Guglielmo. 1993. A null theory of phrasal and compound
stress. LinguisticInquiry24:239-267.
Horvath,Julia. 1986. Focus in the theoryof grammarand the structure
of Hungarian.Dordrecht:Foris.
Horvath,Julia. 2000. Interfacesvs. the computationalsystem in the
syntax of focus. In Interface strategies, ed. by Hans Bennis
andMartinEveraert,183-206. Amsterdam:HollandAcademic
Graphics.
Jackendoff,Ray. 1972. Semantic interpretationin generative grammar. Cambridge,Mass.: MIT Press.
Jacobs, Joachim. 1991. Focus ambiguities. Journal of Semantics 8:
1-36.
Kayne,Richard.1998. Overtvs. covert movement.Syntax 1:128-191.
E. Kiss, Katalin,ed. 1995. Discourse configurationallanguages. Oxford: Oxford University Press.
E. Kiss, Katalin. 1998. Identificationalfocus versus new information
focus. Language 74:245-273.
Koopman,Hilda, and Anna Szabolcsi. 2000. Verbalcomplexes.Cambridge, Mass.: MIT Press.
Krifka,Manfred.1991. A compositionalsemanticsfor multiplefocus
constructions.In Proceedingsof Semanticsand LinguisticTheory (SALT)I, ed. by Steven Moore and Adam ZacharyWyner,
127-158. Cornell Working Papers in Linguistics 10. Ithaca,
N.Y.: Cornell University, CLC Publications.
SQUIBS
AND
DISCUSSION
159
Laka, Itziar. 1990. Negation in syntax. Doctoral dissertation,MIT,
Cambridge,Mass.
Neeleman, Ad, and Tanya Reinhart. 1998. Scramblingand the PFinterface.In Theprojection of arguments:Lexical and compositional factors, ed. by Miriam Butt and Wilhelm Geuder,
309-353. Stanford,Calif.: CSLI Publications.
Neeleman, Ad, and Hans van de Koot. 2002. The configurationalmatrix. LinguisticInquiry33:529-574.
Reinhart,Tanya. 1995. Interfacestrategies.Ms., UtrechtUniversity.
Rizzi, Luigi. 1996. Residual verb second and the Wh Criterion.In
Parametersandfunctional heads, ed. by AdrianaBelletti and
Luigi Rizzi, 63-90. Oxford:Oxford University Press.
Rizzi, Luigi. 1997. The fine structureof the left periphery.In Elements
of grammar:Handbook of generative syntax, ed. by Liliane
Haegeman, 281-337. Dordrecht:Kluwer.
Samek-Lodovici, Vieri. 2002. Prosody-syntax interaction in the
expression of focus. Ms., University College London.
Schwarzschild, Roger. 1999. GIVENness, AvoidF and other constraintson the placementof accent.Natural LanguageSemantics 7:141-177.
Szendroi,Kriszta.2003. A stress-basedapproachto the syntaxof Hungarianfocus. The LinguisticReview 20:37-78.
Williams,Edwin. 1997. Blocking andanaphora.LinguisticInquiry28:
577-628.
Zubizarreta,MariaLuisa. 1998. Prosody,focus, and wordorder.Cambridge, Mass.: MIT Press.
Zubizarreta,Maria Luisa, and Jean-RogerVergnaud. 2000. Phrasal
stress and syntax. Ms., University of SouthernCalifornia,Los
Angeles.
FIRST AND SECOND PERSON
PRONOUNS AS BOUND
VARIABLES
Hotze Rullmann
Universityof Calgary
Dechaine and Wiltschko (2002) argue that, in English, 1st and 2nd
personpronounsbelong to a differentsyntacticcategorythan3rd person pronouns.One of their main argumentsis the claim that English
1st and 2nd personpronounscannotbe used as boundvariables,unlike
3rdpersonpronouns.1In this squib,I discuss datashowing thatEnglish
1st and2nd personpronounsactuallydo allow boundvariableinterpre-
This researchhas been supportedby SSHRCgrant410-2001-1545.1 would
like to thankElizabethRitterand MarthaMcGinnis as well as two anonymous
reviewers for their detailed comments on earlier drafts.
1 Dechaine and Wiltschko propose a new and very interestingtypology
of pronouns supportedwith observations from a wide variety of languages,
arguingthat pronounsfall into three differentcategories:pro-DP, pro-+P, and
pro-NP. They claim that in English, 1st and 2nd person pronouns are proDPs, whereas 3rd person pronounsare pro-4Ps (see Ritter 1995 for a similar
proposal),and since-according to them-DPs are R-expressions,English 1st
and 2nd person pronounscannot function as bound variables.